Welcome to ThrowBack Thursday! Every week we’d like to highlight a blog post from the early days of CosmoQuest. Now, 2012 wasn’t that long ago, but we’ve grown quite a bit since the early days, and we’d like to make sure that everyone gets a chance to read some of the great introductory material that was written when the project launched.

This week, we’re featuring Stuart Robbins’s post from January 10, 2012 titled, “Processing Images for Moon Mappers.” Enjoy!

Introduction

A lot goes into making a picture. Unlike with many consumer cameras today, we can’t just snap a photo and post it online for you to analyze. There are many, many steps and calibrations that go into every single image before we chop it up to serve to you for the different Moon Mappers projects.

Choosing and Downloading Images

The first step is to actually figure out what images we want you to analyze. This selection is driven by the science goals. For the moment, we have selected images around Apollo 15 and a few images on the lunar far side.

Once we know what images we want, we can download them from any number of places. For the moment, our particular favorite is the site put together by the actual camera team, where we search by image ID.

We then download the EDR version of the file. “EDR” stands for “Experiment Data Record” and can be thought of as “straight off the camera.” “CDR” is “Calibrated Data Record” and has had more processing involved; because calibrations (see below!) can change, the EDR is the best to use for our purposes.

ISIS

ISIS is a command-line suite of programs put out by the US Geological Survey. It stands for “Integrated System for Imagers and Spectrometers” and is what pretty much all scientists use to process planetary mission spacecraft data – from Apollo to LRO at the moon to Cassini at Saturn or MESSENGER at Mercury.

The LRO Narrow Angle Camera (NAC) images when downloaded are about 250 MB when downloaded in the standard for the field (IMG format, specifically for the Planetary Data Systems repository (PDS)). The first step of ISIS processing converts the IMG format into an ISIS “CUB” (pronounced “cube”) file. This increases the file size to about 1 GB.

Calibration

Calibration is the next step of processing and this is where a lot of math is done. Every mission team puts out calibration files that tell how every single image has to be corrected from the spacecraft. These include taking into account problems with the data, problems with the detector that make some pixels brighter or darker than others, problems with the actual optics, and how the spacecraft and instrument are pointed.

ISIS uses the calibration files to figure out how to correct every image. But that’s if the calibrations are correct. Early in missions, sometimes the engineers haven’t quite figured out how the calibrations should work. In fact, three years into orbit, they’re still working on how to calibrate the Wide Angle Camera (WAC) color images.

Sometimes, images taken at the same time right next to each other have slightly off calibrations, and this results in weird artifacts. Take images M111578606RE and M111578606LE as shown below, for example. they were taken at the same time, and the right one is properly calibrated, but the left one shows faint black vertical lines.

M111578606RE, Cropped

M111578606LE, Cropped

We actually pointed this case out to the imaging team — this is how science works! Their initial thoughts were that these particular images were taken with the camera not looking straight down – it was slewed to the side – and so that slewing is affecting the calibration.

This is actually one of the images that is being used by Moon Mappers right now, so you may notice some of these vertical stripes. When the imaging team gets back to us and gives us a fix, we’ll upload new images!

Chopping Them Up

Once the images have been fully calibrated (at least based on the current calibration files), the science team sends the output images to the programming team. The programming team chops up the images into little postage stamps, or what we call “slices,” that are 450 pixels on a side. They do this at three different resolutions so that we can get a good range of crater sizes:

1. At full resolution, the images are divided into 450×450 pixel slices. This is done with about 15 pixels overlap on each side to make sure we don’t doom craters at the edge to obscurity.
2. The original NAC image is reduced in size by a factor of 3×, and 450×450 pixel sub-images are then made from that.
3. The NAC strip is reduced to 450 pixels wide (generally a factor of 8-13×) and again divided into 450×450 pixel slices.

Live!

Once this is all done and the programming team does a quick check to make sure things look right, the image slices are uploaded to the server and made live for you to work on.